AI Magazine Volume 23 Number 1 (2002) (© AAAI)
Articles
AI Topics
A Responsibility to
Celebrate AI Responsibly
Bruce G. Buchanan and Jonathan Glick
■ The debut of the AI in the News column elsewhere
in this issue of AI Magazine created a good opportunity to introduce the professional community to
the AI Topics web site,1 home of the AI in the news
virtual page.2 Although AI Topics is designed for
the lay public, it serves a much larger audience.
A
fter a year of planning, the AI Topics
web site went online in 1998 with the
goal of providing students, teachers,
journalists, and others with a basic understanding of what AI is and what AI scientists
do. Much of the genetic code for this web site,
however, was actually written when the American Association for Artificial Intelligence
(AAAI) came into existence. In the 1980 President’s Message delivered by AAAI’s first president, Allen Newell, one finds many of the
responsibilities, principles, and protocols that
have become part of the very fiber of AAAI. Not
surprisingly, some of these elements were
inherited (Newell 1980, p. 3):
[W]e are a society born into a community
of societies. … Because being a scientific
society has grown to be a rather particular
and special thing, it carries with it a rather
special set of obligations—not onerous,
but pleasant, yet nevertheless real obligations.
Scientific societies are for their science.
They are to encourage its happening, to
permit communication about it, to gather
those who want to talk about it in one
place, and in general to conserve, cherish
and celebrate its content.
As a number of factors began to converge in
1997, AAAI started to consider how it should
respond to the public’s growing interest, fascination, and curiosity about AI. No doubt the
increased presence of computers in homes,
schools, and offices was helping to set the stage
by affording a platform on which more and
more people could come face to face with the
steady stream of AI-powered software that was
being made available. Also, friendlier interfaces and free internet access from public
libraries and other institutions were making it
ever easier for the public to go online. Add to
this mix the simplicity and immediacy afforded by e-mail, and one can begin to understand
why the volume of mail from the public to the
AAAI staff and officers was gradually approaching critical mass.
With e-mail as their preferred means of communication, the public displayed an interest in
practically every facet of AI. Although each
and every letter received is unique,3 patterns
did emerge, and it can fairly be said that then,
as now, the questions have tended to touch on
some combination of four general themes:
What is AI: I picked AI as the topic for my
ninth grade science project, and I need to know
who discovered it and how it works. You must
send me the answers as soon as you can because
I have to hand it in at the end of the week.
What has AI accomplished: I’m writing an
article about how technology is making it easier to run a business. Can you tell me about
some actual AI applications? Also, do you
know how much revenue is generated by business software that uses AI?
What are the ethical and social implications of AI: Do you think that a machine can
have a mind and feelings, and if so, does that
mean that they’ll have rights too?
Copyright © 2002, American Association for Artificial Intelligence. All rights reserved. 0738-4602-2002 / $2.00
SPRING 2002
87
Articles
What career options does the field of AI
offer: I’m about to graduate from high school,
and I really like building robots. What should
I study in college and what types of jobs are
there?
The initial wave of questions also gave rise to
one question that never had to be voiced: Why
should AAAI even respond to unsolicited correspondence from the general public? That this
issue was never debated is testimony to AAAI’s
assimilation of the many traditions of scientific
societies that Allen Newell spoke of and that AI
Magazine’s first editor, Alan M. Thompson,
acknowledged in his Welcome to AI Magazine
(Thompson 1980, p. i):
As a major
scientific
society, the
AAAI has a
responsibility
for promoting
its field as
well as
informing its
members of
the latest
technical
developments.
88
AI MAGAZINE
As a major scientific society, the AAAI
has a responsibility for promoting its field
as well as informing its members of the
latest technical developments. … [T]he
editorial committee chose to assign to AI
Magazine the task of providing AAAI
members and the public as well with a
broader perspective on the research activities within AI.
Having now outlined the need and acknowledged the responsibility, the next question
was one of form for, as Newell (1980, p. 2)
remarked, “Given a need, the question seems
always to arise of why this form rather than
some other one.” The simple answer is that we
explored our options and picked the one that
was, to use his words again, “the easiest and
best way possible” (Newell 1980, p. 2). Because
AI Magazine was already serving a different purpose very well, it was neither the simplest nor
the best way to reach the lay public.
With the help of Carol Hamilton, executive
director of AAAI, we explored the possibility of
organizing a network of volunteers who could
respond to the questions arriving at the AAAI
office. What we soon learned was that it was far
easier to enlist volunteers than it was to get
them to promptly respond to the questions we
assigned to them. We then considered updating a general page of frequently asked questions (FAQs) for AI, such as the one that
Carnegie Mellon University was then offering,
but decided against recruiting volunteers to
rewrite and maintain answers. Although we
certainly were aware of the material written by
members of the AI community and published
for a wide audience (for example, McCarthy
[1966]; Duda and Shortliffe [1983]; and Feigenbaum, McCorduck, and Nii [1988]), there were
not enough of them, and much of the material
was only available in large research libraries.
Another idea we investigated involved creating
a variety of teaching modules for science teachers to make it easier for them to bring AI topics
into the classroom, but we were timid about
moving into K–12 curriculum design and overcoming the tight constraints on teachers’ use
of classroom time. Moreover, these modules, as
well as practically every other approach we
considered, would require many people to
make commitments to assume the never-ending burden of keeping everything current.
However, it was during the fact-finding forays and benchmarking exercises for these and
the other options we explored that we became
ever more impressed with the amount of good
material pertaining to AI that could freely be
accessed on the web. Such resources came in all
shapes and sizes and included virtual encyclopedias and magazines, what-we-do pages posted by AI labs, student projects, descriptions of
commercial products, and materials posted for
university courses. Soon we came to realize that
there were enough really good resources out
there to serve as the foundation on which a
web site could be built. Our standard for determining what constituted a really good resource
was Scientific American, a magazine that represented our paradigm of clear exposition. Part of
the secret for its effectiveness was mentioned in
a review of Gerard Piel’s new book, The Age of
Science: What Scientists Learned in the 20th Century (Basic Books, 2001) (Graham 2001):
The distinguishing characteristics of
Piel’s science journalism were a conviction
that science is both objective and a social
good, a refusal to talk down to his audience and careful, even lavish illustrations
and photographs. His ideal reader was a
scientist seeking knowledge outside his or
her own field. He often enlisted scientists
as authors, coaxing them with meticulous
editing to find ways of explaining their
recent work to intelligent people with other backgrounds. A nonscientist (he was an
undergraduate in history at Harvard). . .
When Piel and his colleagues created
Scientific American over a half-century
ago, he correctly saw the need for a magazine that would transcend the specialist
science journals and offer overviews for
educated lay people of developments in
the individual scientific disciplines.
It is worth pausing at this juncture to
acknowledge the importance of making understandable scientific information available to
the lay public and how it eclipses both the science of AI and the present moment. In Closing
the Knowledge Gap between Scientist and Nonscientist, Takashi Tachiban’s 1998 contribution to
the Science series, Essays on Science and Society, he looked back over almost 40 years (Tach-
Articles
iban 1998, 778–779) and observed.
In his classic work Two Cultures and the
Scientific Revolution [(Cambridge University Press, 1959), originally a Rede Lecture
presented at Cambridge University (1959)],
C. P. Snow wrote that the chasm between
scientists and literary intellectuals was so
vast that they could not communicate
with each other. For example in physics,
Snow found the mid–20th century literati
to be as ignorant about science as had been
their Neolithic ancestors. And this situation that Snow lamented nearly four
decades ago has only grown worse. The
current level of basic scientific knowledge
is so low that it is difficult to interest even
the brightest layman or nonscience student in what modern science is doing. I
border on despair at my inability to keep
them interested long enough to correctly
understand both a specific research project,
and its aims. The chasm between scientist
and nonscientist has widened to become a
gulf. And it is the task of science and society to narrow that gulf through an intellectual shift of tectonic plates.
Modern scientists can no longer expect
to live their lives in proud isolation
because most significant scientific
research requires substantial funding. …
C. P. Snow was talking mainly about
intellectuals. But, today the reactions of
ordinary people and the mass media matter more. Since politicians are easily
swayed by their perceptions of public
opinion, key to promoting wise political
decisions about scientific matters is a
sound understanding of science among
the general population and the media that
feed, reinforce, and mobilize its views.
Further evidence that these issues have not
gone away can be found in the House Committee on Science’s 1998 Report to Congress:
Unlocking Our Future—Toward a New National Science Policy.4
One of the ironies of our modern age is
that although our society depends on science as never before, what scientists do
remains an enigma to most people. As any
nonscientist who has tried to wade
through a scientific journal knows, the language of science is virtually incomprehensible to the layman. While these journals
are not written for a general audience—nor
should they be—they are perhaps the clearest example of the widening chasm
between scientists and the rest of society.
If we are to maintain public appreciation and support for our scientific enterprise, a way to translate the benefits and
grandeur of science into the language of
ordinary people is sorely needed. Scientists have wonderful stories to tell, yet too
often they get told poorly, if at all. Educators and journalists have a role to play in
communicating the achievements of science, but scientists must recognize that
they, too, have a responsibility to increase
the availability and salience of science to
the public.
The online availability of understandable
resources seemed to address the question of
appropriate form. However, it raised two new
questions for us: Could a web site fulfill our
responsibilities, and would it be acceptable to
build the site with materials that AAAI did not
author or otherwise create? Once again,
guidance could be found in Newell’s President’s message (1980, p. 3):5
The first watchword is cooperation—
cooperation with all the others that are
trying to nurture artificial intelligence… .
As a second aspect of style, we will
innovate—not for the sake of innovation
(I trust), but because problems need to be
solved in the easiest and best way possible. For instance, we hope to find ways to
exploit the computer and computer networks. We could hardly expect less of
ourselves.
Thus freed from pressures that might have
led to creating all new text and excited by the
potential of a web site, we shifted our focus
from creating resources to creating a means to
get the right resource to the right person.
Underlying this approach was the realization
that although the vast universe of online information contained many stellar resources, there
was considerable redundancy and misinformation, not to mention space junk and the occasional flaming comet. Moreover, the volume of
material presented substantial problems for
laypersons who just wanted to read concise,
accurate descriptions. The decision was then
made to mediate the relationship between the
information seeker and the information by
developing a safe and comfortable environment where what was sought could simply be
found. Making it a reality entailed (1) identifying readable overview articles that are readily
available (almost all on the web) about major
topics and some of the specialized topics that
are mentioned in the popular press; (2) organizing the information in a structure that is meaningful to nonprofessionals, with ample cross-
… freed from
pressures that
might have
led to creating
all new text
and excited
by the
potential of a
web site, we
shifted our
focus from
creating
resources to
creating a
means to get
the right
resource to the
right person.
SPRING 2002 89
Articles
Figure 1. AI Topics Home Page (Top Portion).
referencing and redundancy in the indexing;
(3) providing additional items that make AI fun
and interesting and encourage further study;
(4) maintaining contact with the users by both
direct means (inquiries, feedback) and indirect
means (using data from the server) to assure
that the site remains responsive to their needs;
and (5) finding someone to do it!
Actually, it wasn’t just someone—it was
someone to do it all. Locating and sorting
through the understandable information that
could be found online in diverse forms and in
obscure places was only the beginning
because it then had to be packaged so that it
could effectively be utilized. In addition, it
was this need to objectively organize the
material for a lay audience that tipped the
scale in favor of finding a trained librarian
who, unfettered by a technical background,
could bring a fresh approach to building and
structuring the collection. The other side of
this coin, however, is that this individual
would have some scientific naïveté; so, it was
critical that scientists oversee the effort
90
AI MAGAZINE
because they are sensitive to the specialized
terminology and flame issues in the field.
Without boundaries, one person could not
be expected to organize the entire universe of
introductory AI resources online for all segments of the lay public. Thus, to help with the
process of staking out borders, three decisions
were made: (1) the primary target audience
would be English-speaking high school students from around the world; (2) material
would come from stable sources that are available primarily online without charge; and (3)
use and enjoyment of the site should not be
dependent on high-speed connections, the
newest hardware, or the latest software.
High school students were the ideal primary
audience because they are the future of AI and
have adapted well to the web. High school
teachers and journalists were selected as the
secondary audience because they nurture the
interests of the primary audience. Notwithstanding having defined these focus groups,
the web site serves a much broader audience of
students and inquisitive laypersons that
Articles
includes undergraduates and graduate students
not in AI as well as the parents and even the
younger siblings of high school students. However, the fact remains that although one can
find a page of resources collected especially for
older students in the site,6 and another special
collection for educators,7 there is no single
page expressly dedicated to parents.
Also, the decisions to focus on material
available on the web and to keep the site “lowtech” were made with the primary audience in
mind. It does little good to point students to
texts that can only be found in a local library
when they are often without access to such a
facility.8 Likewise, because it cannot be assumed that students who come to the site from
around the globe will have the latest computer
equipment or an optimum internet connection, we keep the site itself simple and basic
(although we do point to video clips, interactive demos, and the like).
Another way that we defined boundaries was
to introduce our springboard approach. At its
heart, it is the recognition and honest
acknowledgment of (1) what we can reasonably accomplish; (2) what can reasonably be
expected from any web site and, specifically,
this web site; and (3) what we can reasonably
expect from our visitors. As we explain on the
Springboard—How to Use This Site page:9
AI Topics seeks to provide information
that can be understood by those who are
beginning to learn about artificial intelligence. Because much of this audience has
little or no technical background, the
information must be general in nature and
will often come from non-scientific publications. We have used our best efforts to
strike a balance between scientific exactitude and a presentation that is not daunting. Furthermore, AI Topics is a resource to
be utilized as a starting point, or springboard, rather than as a collection of blackletter dogma. In fact, given that the field of
artificial intelligence is diverse, dynamic,
and evolving, we want you to understand
that we simply cannot stay current with all
of the developments and resources, nor are
we able to represent all viewpoints… .
AI Topics is a springboard in the sense
that it can launch you on your way to
more resources. It is our hope that by providing some basic information, we can
equip you to further explore AI in more
depth on your own. By using keywords,
project names, articles and authors, and
other terms found in these pages, you can
frame a fruitful search request using any
of the popular search engines as well as
various online databases
library’s online catalog.…
and
your
… With one webmaster to maintain more
than 100 web pages, there is simply no
way that every page and every topic can
be kept 100% current. From the pedagogical perspective, we don’t want to give the
impression that our resource base is functionally complete lest someone infer that
the presence or absence of a particular
resource is of significance. [This is also
covered on our page of Notices along with
other related information.]
This approach is consistent with our decision to present material about AI fairly, but
favorably. In particular, we chose to celebrate
AI’s numerous definitions and approaches
rather than present a false picture of a unified
whole, present the state of the art as an
achievement as well as an invitation to imagine what awaits over the horizon, welcome differing viewpoints as possibilities rather than as
conflicting positions, recognize responsible
criticism as a mirror for self-examination, and
acknowledge that multiple access paths and
index terms are not necessarily duplicative.
How these principles and guidelines become
woven into the fabric of the AI Topics site can
be illustrated by looking at the site from the
perspective of the virtual AI in the News
page.10 This page consists of a collection of
news nuggets distilled from recent news articles appearing in the popular media and consist of an excerpt from the article, its URL, and
pointers to related pages within AI Topics. A
typical, though in this case fictional, nugget
would look something like:
November ##, 2001: Face-Recognition
To The Rescue? By Arthur Paige. XYZ News
Bureau / available from Ourtown Gazette.
“Using artificial intelligence technology
similar to what enables a robot to recognize an image from different angles and
under various conditions, the ABC Company developed a system that can be used in
airports to screen passengers…. Many people are concerned that such systems, as
well as similar systems that use voicerecognition or fingerprints … are unreliable and pose a threat to our civil liberties.”
Without
boundaries,
one person
could not be
expected to
organize the
entire universe of
introductory
AI resources
online for all
segments of
the lay public.
>>>Image Understanding, Vision, Robots,
Speech, Pattern Recognition, Ethics and
Social Implications, Law Enforcement,
and our biometric toon.
Initially, a distinction must be made between
the nugget and the article from which it was
taken because each has a role independent of
the other. Once we’ve captured a nugget, we
SPRING 2002 91
Articles
AI in the News
AI Overview
Agents
Applications
Cognitive Science
Education
Ethics and Social Implications
Expert Systems
FAQs
Games and Puzzles
History
Interfaces
Machine Learning
Natural Language
Philosophy
Reasoning
Reference Shelf
Representation
Resources
Robots
Science Fiction
Speech
Systems and Languages
Turing Test
Vision
Special Pages
Table 1. The Current Set of Topics.
Note: Exploring the site will give you a much better sense of its scope than lists
can. We do not list all the 100+ subtopic pages underneath these in the AI Topics
site or show the hundreds of index entries.
Another way
that we
encourage
further study
is by not
offering the
answer.
92
AI MAGAZINE
can reliably integrate it into the site, unlike the
article itself that we have no control over and
that might soon migrate into the publisher’s
fee-based archive or even vanish.11 Nevertheless, during their accessible lifetimes. These
items typically provide a wonderful source of
basic, understandable information and often
include exciting illustrations, charts, and photographs. However, as we explain, even when
the link that we provide to the article has died,
the nugget lives on because we almost never
remove an item from the web site. In the case
of news stories, once an item has been posted
on the news page for a month or two, it is
moved to our permanent news archive collection. In addition, rather than simply delete other resources that have appeared within AI
Topics, we warehouse them in our Page Archive
or cemetery of Dead Links.
Standing alone, the nuggets provide a stage
from which we can launch a number of initiatives. First, we can use them to introduce the
site’s topics (see table 1 for the current set of
topics) and subtopics in the context of something concrete and not intimidating. In contrast to our site map and index that list what
for many are simply strings of abstract concepts, the news page allows the visitor to make
a connection between something familiar and
the related page(s) within the site.
On another level, the collection of nuggets
serves as evidence that the science of AI is alive
and well and contributing to actual, real-world
applications—even if some are controversial.
However, this, too, is an important function of
the news page. Allen Newell recognized that AI
like other sciences, “touch[es] on the place and
nature of us humans within the universe,” and
he “believe[d] that the controversies have their
natural cause in the type of knowledge our science reveals” (Newell 1980, p. 4). He cautioned,
however, that “they must be dealt with on the
basis of substance and truth” (Newell 1980, p.
4). Thus, provided that the source is responsible, reference to such articles is part of a healthy
dialogue about important issues engendered by
our science. Moreover, we place the issues within a larger context by pointing visitors to the
Ethical and Social Implications page and
overviews of relevant technical topics.
Viewed holistically, the string of nuggets on
the news page conveys a sense of momentum
and excitement that reinforces AI’s vibrancy.
Also contributing to this effect is that the page
is updated several times a week.12 Thus, one
can read articles foreshadowing the competition for the Loebner Prize and then return to
the site a few days later to find articles about
the outcome. Sometimes the fact that two
events have become linked in this string of
nuggets results in a very powerful experience,
as was the case when articles about the
RoboCup rescue competition held in Seattle in
August 2001 were soon followed by articles
about some of these same competitors (both
human and robotic) arriving at Ground Zero in
New York City to lend assistance in the wake of
the tragedy of September 11, 2001.
The ability to add nuggets to the news page
several times a week, or sometimes several
times a day, is just one of the ways that AI Topics can quickly accommodate current events.
AI Topics can also easily be modified to address
changing circumstances. The field of biometrics offers a good case because although it has
been around for quite some time, the events of
September 11, 2001, propelled it into the spotlight. Because the term biometrics began to
Articles
appear more and more often in news
articles mentioning AI techniques, we
added the term to the site’s A–Z Index
with pointers to related pages and a
cross-reference to the index term surveillance. We also created a special collection of resources about biometrics
that we placed on the Image-Understanding page because most of the
boundary lines in AI Topics are drawn
in the sand and not in cement, we
have the flexibility to cut and paste
the resources to the Pattern-Recognition page if such repositioning would
be justified at some future time. To
some extent, however, it is a moot
point because the relevant pages are
linked by pointers among themselves
so that someone reading about biometrics on the Law Enforcement page
can follow the pointer to the ImageUnderstanding page, home of the biometric collection that contains links to
Pattern Recognition, Law Enforcement, and other pages.
It is also quite common for us to
post several pointers to the same news
story throughout the site. Thus, one
can follow a link from a news nugget
to the Law Enforcement page only to
find the same nugget that provided
the link to that page. We do this multiple posting because people can arrive
at the Law Enforcement page by any
number of paths, many of which
might not include a stop at AI in the
News. Although the circularity, repetition, and cross-referencing might be
redundant, it does increase the
chances of people finding what
they’re looking for as well as offering
opportunities to explore tangents.
Because redundancy is a concept
worth revisiting, it should not go
unnoticed that much of the redundancy within the site is a direct result
of indirect feedback. As discussed earlier, inquiries from the public were an
impetus for creating AI Topics. These
questions continue to be an umbilical
cord to our audience and enable us to
keep the site fine tuned.13 When, for
example, we receive an inquiry asking
about AI applications in a specific field
such as law enforcement, two issues
are immediately raised: (1) do we currently offer resources about this subject and, if not, should we and (2) if we
already offer relevant resources, why
couldn’t the person find them. This
exercise has resulted in countless additions to the A–Z Index, lots of crossreferencing, and even several new
pages of subtopics.
Because our list of objectives also
includes making AI fun and interesting, we obtained the humorist’s version of the poetic license and began to
create cartoons. One of these cartoons
is our biometric toon that plays off the
ability of AI-powered biometric systems to identify people by their gait,
voice, or appearance. The cartoon,
Biometrics 101,14 is posted on our
toons page and depicts a duck prancing adjacent to a block of text that
reads: “If it walks like a duck and
sounds like a duck and looks like a
duck, then it must be a duck.” Beneath
the cartoon is a brief description of
biometrics extracted from, and linked
to, a good overview article, all of
which is followed by a link to our collection of resources addressing the
intersection of AI and biometrics. In
this way, we hope to take the interest
created by the cartoon and channel it
to resources for further study.
Like the cartoon, the format of the
AI in the News page is structured to
encourage further study. Seeking to
build on the energy created by a person’s interest in any given news
nugget, we entice this person to continue exploring the main topic of the
article or other subjects raised therein
by providing links to pages within AI
Topics and the article itself. The captivating nature of the specific examples
makes AI in the News the perfect centerpiece of our response to any student
who sends an inquiry along the lines
of, “I picked AI as the topic of my term
paper and I need to find something to
write about.” Not only does the news
page offer the student a parade of
ideas to choose from, but each nugget
includes everything he/she needs to
get started.
Another way that we encourage further study is by not offering the
answer. Although we are more than
happy to help students, we will not do
their work for them. Also, most questions do not have the simple, definitive answer that many students are
looking for. Moreover, were we to take
a position regarding, for example, AI’s
single greatest achievement, which
seems to be a popular topic for school
reports, the site would appear to be, or
possibly become, a bully pulpit. However, our recognition of, and respect
for, the fact that both AAAI, as well as
the science of AI, embraces a diverse
collection of viewpoints results in a
nonjudgmental stance that fits neatly
with this egalitarian principle and the
springboard approach.
AI in the News has also encouraged
us to embark on some new initiatives.
For example, when it became apparent
that the movie A.I. was generating a
steady stream of prerelease media coverage about the movie and the science,
we created the AI: The Movie page.15
The page serves as both a special substantive portal into AI Topics and a
place where we can assemble and
organize a collection of news items
occasioned by this well-publicized
event. As evidence that the page was
effective, the movie page is featured
on the official Warner Brothers A.I.
web site;16 the AI Topics site received a
glowing review (Leslie 2001) in Science
that specifically referenced the movie
page just one week before the movie’s
June 29, 2001, release in the United
States; and the AI Topics site administrator reported that the movie page
received approximately 3380 hits in
June 2001 and 3500 hits in July 2001.
The news page has also propelled us
into two new initiatives outside the
web site. The first is the electronic
newsletter, AI Alert, which is now distributed approximately twice every
month to AAAI members. AI Alert
consists of an eclectic subset culled
from the virtual AI in the News page
and provides the professional community with a glimpse of AI as seen from
the outside looking in. AI Alert consists of an eclectic subset culled from
the virtual AI in the News page and
provides the professional community
with a glimpse of AI as seen from the
outside looking in. The periodic view
from this perch provides not only a
reminder that one’s family and
friends, as well as potential funders
and future students, might see AI in
quite a different light but also an invitation of sorts. As explained by Norman Augustine in his article, “What
We Don’t Know Does Hurt Us—How
SPRING 2002 93
Articles
Scientific Illiteracy Hobbles Society, for
the Science series, Essays on Science
and Society (Augustine 1998, p.
1640–1641):
To a not inconsiderable segment
of the public, the word “science”
conjures up images of Chernobyl,
Bhopal, Thalidomide, Challenger, and the atomic bomb. Too
often science is perceived as the
cause of problems rather than the
solution, as something to be
avoided rather than something to
be embraced.
As part of his solution, he suggests
(Augustine 1998, p. 1641):
[L]iving as we do in a “sound
bite” world, scientists must learn
to communicate far more effectively with nonscientist audiences. In my judgment, this
remains the greatest shortcoming
of most scientists and engineers
today. The time has arrived when
scientists will have to come down
from the Ivory Tower and enter
the arena of real-world debate,
bubbling controversy, and—brace
yourselves—politics.
The other initiative we’ve been
encouraged to explore involves giving
the news nuggets a presence in the
print medium. Our first foray appears
elsewhere in this issue of AI Magazine
as the AI in the News column. In contrast to the e-mailed AI Alert, the magazine column provides a more permanent, although much less extensive,
record. However, in spite of its size, we
expect that this page of mininuggets
has much to offer those who read it
now as well as those who will come
upon it in 20 years.
The AI Topics site is intended to be
a window into the science of AI and to
partially address AAAI’s responsibility
to explain what we do. Your suggestions, as well as any resources that
you’d like to contribute, are important
to its success. Even more important is
your participation in ongoing discussions about AI, for which we hope the
site will provide useful input.
Acknowledgments
AI Topics could neither have been
launched nor have survived this long
without the ongoing help and support
94
AI MAGAZINE
provided by Carol Hamilton, Mike
Hamilton, Richard Skalsky, and others. We also owe a debt of gratitude to
Karen Peterson, who, as the web site’s
first librarian and developer, helped
establish the directions and standards
for the site.
Notes
1. www.aaai.org/aitopics.
2. Some of the material in this article previously appeared in Buchanan, B. G., and
Glick, J. 2001. AI Topics—Organizing
Online Knowledge Sources about AI for the
Lay Public. ACM Intelligence 12(3): 25–30.
3. A selection of actual e-mails we’ve
received can be found on the Mailbox page
at www.aaai.org/aitopics/html/mailbox.
html.
4. House Committee on Science. 1998. A
Report to Congress: Unlocking Our
Future—Toward a New National Science
Policy. Preface to Chapter V (Sustaining the
Research Enterprise—the Importance of
Education) Part C (Communicating Science). Available at www.access.gpo.gov/
congress/house/science/cp105-b/science105b,pdf or www.house.gov/science/
science_policy_report.htm.
5. Newell’s use of the term style relates to
“How will the AAAI conduct itself?” and
“What will be its attitudes toward doing
this or that?” (Newell 1980, p. 3).
6. Resources for students. www.aaai.
org/aitopics/html/ resources. html.
7. Resources for educators. www.aaai.org/
aitopics/html/resourcED.html.
8. We use the word access in its broadest
sense to include both the predicament of
needing something at 2 AM and the library
is closed and the situation that “M,” an
eighth-grade student, described for us in an
e-mail message:
Hello ,
I am doing a science project on a topic of
my choice and I chose to do AI, which is
something I am very interested in and I was
just wondering if you could send me some
information about it or maybe give me
some webpages to go to. You see I live in
Yellowknife, N.T, Canada and so I appart
[sic] from the web I am very restricted to
resources. If you could send my [sic] some
info. it would be a great help. Please Reply.
9. www.aaai.org/aitopics/html/springbd.
html.
10. www.aaai.org/aitopics/html/current.
html.
11. However, because the same article
might appear in several publications, both
online and in print, a person can use the
citation we provide to try to track down its
other incarnations. It is also not uncommon for a reprint of the article to find its
way onto the web site of the lab or company featured in the article.
12. In an unsolicited e-mail message, “J”
told us: “Just wanted to let you know that I
check the “news” section of your website
frequently and I appreciate how current
you keep it. Excellent work.”
13. In the bottom corner of every page in
AI Topics is a link to the Inquiries—Feedback—Contact Us page to remind the audience that we’re only an e-mail away.
14. www.aaai.org/aitopics/html/toons.ht
ml #biomet.
15. www.aaai.org/aitopics//html/aimovie.
html.
16. aimovie.warnerbros.com.
References
Augustine, N. 1998. What We Don’t Know
Does Hurt Us. How Scientific Illiteracy
Hobbles Society. Science 279(5357):
1640–1641.
Duda, R. O., and Shortliffe, E. H. 1983.
Expert Systems Research. Science 220(4594):
261–268.
Feigenbaum, E. A.; McCorduck, P.; and Nii,
H. P. 1988. The Rise of the Expert Company.
New York: Random House.
Gerard, P. 2001. The New York Times Book
Review, 11 November, 29.
Graham, L. 2001. Review of The Age of Science: What Scientists Learned in the 20th Century, by Gerard Piel, New York Times Book
Review, 11 November, p. 29.
Leslie, M., ed. 2001. NetWatch. Science
292(5525): 2223.
McCarthy, J. 1966. Information. Scientific
American 215(3): 64-73.
Newell, A. 1980. AAAI President’s Message.
AI Magazine 1(1): 1–4.
Tachibana, T. 1998. Closing the Knowledge
Gap between Scientist and Nonscientist.
Science 281(5378): 778–779.
Thompson, A. M. 1980. Welcome to AI
Magazine. AI Magazine 1(1): i.
Bruce Buchanan (M.S.,
Ph.D.) is the immediate
past-president of the
American Association for
Artificial Intelligence,
university professor of
computer science at the
University of Pittsburgh,
and a founder of the AI
Topics web site. His e-mail address is
buchanan@cs.pitt.edu.
Jon Glick (B.A., J.D., M.L.I.S.) is the current
web master of the AI Topics web site. His email address is aitopics@aaai.org.